FAQ: Genome-Phenome Analysis Platform (FAQ)

The GPAP allows the clinicians and researchers who submit patient datasets to analyse their data themselves. You remain in control of the analysis of the data you submit – it is not analysed by others on your behalf, although you can get a lot of support from the RD-Connect bioinformatics specialists to help you make the most of the system. If this is the first opportunity you have had to analyse your data, you can request a 6-month embargo period so that your group has the chance to take a look before others can access it.

The GPAP is also a mechanism for data sharing – allowing you to access information about similar patients submitted by other users and enabling other users to query the datasets you have submitted. This is an essential feature for gene discovery work, allowing matchmaking, finding second families, finding patient cohorts for validation studies, and allowing basic science researchers to find human analogues for their animal models. Sharing data benefits participating researchers through greater numbers of publications and new collaborations.

RD-Connect GPAP has been designed for rare disease clinicians and researchers around the world, particularly those who do not have extensive in-house bioinformatics support. This includes clinicians from European Reference Networks and researchers from multiple national and European research projects.

Yes! If you are interested in finding human data analogues for your gene or model of interest or information about whether a specific variant you are working on in an animal model causes a similar phenotypic presentation in a human, please consider registering. You will have to state your research interests during the signup process but we do not require you to submit data.

The GPAP is available free of charge to researchers and clinicians validated in the process established by our ethical-legal experts. In the case of research groups, the PI needs to be verified first and then he/she can enrol his/her other group members, provided the PI has the authority to “vouch for” the users in his/her group. User registration and validation involves three steps. First, you need to register at [url]. We’ll ask you to provide proof of your credentials as legitimate researchers, a scan of your ID (e.g. passport), and explain your research interests, so we can verify that you are a researcher or clinician. You will also need to read and accept the RD-Connect Code of Conduct by signing an Adherence Agreement (example here). Your application will be considered by the Data Access Committee (DAC) and you will receive a response within 10 working days.

Second, you will need to provide the phenotypic information about each of your patients (min 5 Human Phenotype Ontology terms per patient) in the RD-Connect PhenoTips instance. This step is necessary before you can upload the genomic data of the patients in question.

Third, you will transfer a copy of your raw sequencing data. Once they are securely stored and processed through the RD-Connect pipeline, you will be able to analyse them in the GPAP. Validated users are granted access for 12 months and their use of the system is monitored to ensure that it is in line with their stated research interests.

Go to Genome-Phenome Analysis Platform Registration page.

In short: you need to ensure you have patient consent, and you need to ensure you have the data in the correct format (see below). Privacy and safety of patient data is secured in line with the GDPR and local permissions, and patients’ identifiable information (names, date of birth, address) must NOT be submitted. This includes “incidental” ways in which patients might be identified, for example through the use of initials in filenames: if we suspect that a filename might contain such data, we will not allow its submission. You need to check whether you have obtained consent from the patient to share data with other databases in Europe for rare disease research. You need to check whether you have the phenotypic (HPO) and the genomic (FASTQ or BAM) datasets in the correct formats.

We accept human genomic, transcriptomic and phenotypic data meeting the following requirements:

• Each file or dataset must be about an individual person, not an aggregate dataset that contains data from multiple individuals.
• We only accept datasets on individuals with a rare disease or their family members (affected or unaffected).
• We accept both solved and unsolved cases.
• We accept whole-exome, whole-genome and gene panel sequencing data.
• You must be able to send us the “raw data” from the sequencing experiment (FASTQ or untrimmed BAM format).
• Each genomic or transcriptomic dataset must be paired with detailed clinical information (information about the patient’s phenotype). We ask you to enter this in PhenoTips, an online interface that allows it to be converted into Human Phenotype Ontology (HPO) terms. See our video tutorial and PhenoTips guide.

One dataset is considered to include both phenotypic data in the form of human phenotype ontology (HPO) terms and exome/genome data of a patient.

There are two options for uploading datasets.

• Data collections from centres that will upload fewer than 100 datasets: datasets will be uploaded using the standard RD-Connect GPAP upload interface, which includes user friendly PhenoTips templates to enter the phenotypic information, and user-friendly tables to upload the genomic data and metadata. An option to bulk upload the genomic data and metadata is also available.
• Data collection from centres that will upload 100 or more datasets: please discuss with CNAG-CRG in Barcelona ([email protected]) a customised bulk upload option to facilitate uploading large data collections.

All sequencing data is submitted as raw data in FASTQ (or BAM) format and is processed through the standardised pipeline. For more information, see the section “Data processing and storage” below.

All phenotypic data are submitted as HPO terms. Diseases are classified using OMIM/ORDO. Both phenotype and disease classification are entered in the RD-Connect PhenoTips instance, which makes it simple for clinicians to code their data in this manner. Clinical interpretation of the data (final or temporary) can be entered in the GPAP but this is not a requirement at upload. See how to upload data in PhenoTips >>

Information on kinship is collected by the clinician and submitted as part of the submission to the PhenoTips database. For each family member, next generation sequencing data and PhenoTips forms are submitted individually to the database. Samples can then be linked into Families, using a built-in pedigree drawing tool in PhenoTips. Multiple samples from the same family can be analysed in the GPAP simultaneously.

Once you upload your data to GPAP, the following steps happen:

1. Processing. Data is put through a standard analysis pipeline involving alignment, variant calling and annotation, resulting in a .bam file containing all read data aligned to the reference genome, and a .gvcf file containing the called variants, the positions called as reference, and annotations from the pipeline.

2. Embargoed availability (availability to the submitter’s group alone). As soon as we have processed the data you have submitted, we will send you an email to let you know it is ready for you to view in the GPAP. When the processing is complete, two things happen:

• The original submitted raw data file (.fastq or .bam) and the .bam file produced from the RD-Connect standard analysis pipeline are submitted to the European Genome-phenome Archive (EGA), for long-term storage on behalf of the original submitter and under the submitter’s ownership.
• The processed data is made available to the submitter through the GPAP for analysis by the submitter’s group. Ownership is assigned to the group of the submitting PI. If you have requested an embargo period for the data you submitted, then at this point, datasets at EGA are “invisible” in the EGA system and not requestable by external users, and datasets in the GPAP are only available to members of your group. This is now the time during which your group can access the data and do your primary analysis.

3. Controlled-access availability. After the expiry of the embargo period, two things happen:

• The datasets submitted to the EGA become “visible” in the EGA catalogue, allowing EGA users to know they exist. If an EGA user wishes to access your dataset, they must request access via the standard EGA request mechanism, meaning the data access committee you nominated will approve or deny every request.
• The corresponding datasets in the GPAP are now accessible to other authorised users for queries. During this time, you can see which other users have queried your data. You may receive requests for collaborations if other researchers find a variant or phenotype of interest in your data, in the same way that you may already have contacted other users if you found something of interest in their data when comparing it with your own.

When you submit your raw genomic data file and the corresponding phenotypic data, it is processed on servers at CNAG-CRG in Barcelona by bioinformatics team members there. All sequencing data are submitted as raw data in FASTQ (or BAM) format and are processed through the same standard pipeline. This ensures that data from different sequencing providers are comparable. The processed data are accessible online for real-time analysis in the GPAP.

The raw genomic data are stored for long-term access at the European Genome-phenome Archive (EGA), a secure, controlled-access repository. The EGA serves as an archive for publications as well as data on several levels from case control, population, and family studies. This includes raw data, which allows future reanalysis using other algorithms and genotype calls – information about pathogenic genetic variants, such as single nucleotide variants (SNVs) and copy number variants (CNVs) provided by the data submitters.

The EGA provides the necessary security required to control access to the data and maintain patient confidentiality. Data can be accessed only by authorised researchers and clinicians. In all cases, data access decisions are made not by the EGA but by an appropriate Data Access Committee, which can be the person or group submitting the data.

At the time of uploading a dataset to the GPAP, the user can indicate if the dataset is already available at the EGA and provide the corresponding reference number. For datasets not yet available at the EGA, the CNAG-CRG will broker the submission to the EGA of the data and metadata uploaded to the GPAP. The original data submitter will be responsible for making decisions regarding the future access to their datasets.

The processed data (the called variants, not the raw data) is securely stored on the GPAP servers at CNAG-CRG in Barcelona.

Yes! One of the central principles of RD-Connect is making data FAIR: Findable, Accessible, Interoperable and Reuseable. All genomic datasets are reprocessed through a standard pipeline which results in a standard gVCF format file per sample that is not only uploaded into the GPAP for analysis but can also be made available as a download to the submitter on request if they should wish to take it elsewhere in future. Phenotypic data are submitted to PhenoTips in the form of Human Phenotype Ontology (HPO) terms, which ensures standardization and computer-readability of the phenotype and thus facilitates any future transfer to other computer systems making use of the HPO, which is the most widely used ontology for rare disease phenotypes and is recommended by the International Rare Diseases Research Consortium, IRDiRC.

 

 

Short answer: as the submitter, you do!

Longer answer: The primary goal of GPAP is to give you, as a rare disease researcher, the tools to better analyse your own patients: control over the primary data analysis remains with you. In addition, the GPAP enables the data you submit to be shared with other researchers who have access to the system. If you wish, you can request a 6-month embargo period, when you and your team have exclusive access to the data. After this period, it becomes accessible to other users. This allows you to find confirmatory cases or second families for your interesting candidate genes, but it also allows groups who may have completely different research questions to benefit from the availability of your data.

From time to time, the GPAP development team in Barcelona may run some analyses on your data to test the tools that we incorporate into the system and in some cases to provide you with interesting results or candidates that you might wish to follow up yourselves. For example, with our runs of homozygosity tool, we were able to point out to some submitters that some of their cases were consanguineous when this was not known at the time of submission – this allowed some interesting candidates to be found in homozygous regions. In cases where we do this, we provide the interesting results to you as the submitter so that you can follow them up – we would not publish gene discovery or case report papers on your patients ourselves. We do have an interest in publishing aggregate data and in methodology and statistical papers that show the value of the system. In these cases, we credit all data submitters with an acknowledgement.

Clinical interpretation of the data (final or temporary) can be entered in the GPAP but this is not a requirement at upload.

A number of inbuilt functionalities allow identifying disease-causing genetic variants in patients and solving even difficult cases.

Basic filters Users can analyse one or multiple individuals, e.g. patient and family members. The results can be filtered by quality, mode of inheritance (e.g. autosomal, homozygous recessive, compound heterozygous, etc.), control population frequencies (e.g. 1000 Genomes Project, gnomAD)  and known and expected pathogenicity according to ClinVar database and several predictors.

Runs of Homozygosity (RoH) allow identifying consanguineous cases even when not identified as such by the clinician. RoH narrows the search down to the homozygous regions in the patient’s genome, which are more likely to be contain the disease-causing variant.

Phenotypic data The Platform matches individual’s sequencing data with detailed clinical information about his/her symptoms (deep phenotyping), which are stored in the Human Phenotype Ontology (HPO) format in the PhenoTips database. The Platform allows refining the genomic analysis results by selected HPO-encoded symptoms, or genes related to a specific disease in the OMIM (Online Mendelian Inheritance in Man) database.

Built-in Exomiser The fully-integrated Exomiser tool extracts clinical information from PhenoTips automatically and highlights the candidate variants best matching the patient’s symptoms.

Patient matchmaking allows finding individuals with variants affecting the gene of interest. Presence of similar symptoms in the matching patients is a strong hint that a given variant is disease causing and helps confirm genetic diagnosis. Matchmaking can also help basic researchers learn how mutations in the gene they study affect humans.

Connection to other databases To make gene discovery easier, all variants on the results list have direct links to variant descriptions in external databases, such as OMIM, PubMed and Ensembl.

RD-Connect bioinformatic tools In addition, RD-Connect has developed several bioinformatic tools to help researchers analyse omics data and identify targets for potential therapies. Read more >>

To learn how to perform those analyses, watch our GPAP video tutorials.

GPAP is integrated in the Beacon Network (https://beacon-network.org), a project by the Global Alliance for Genomics and Health (GA4GH).

GPAP participates in MatchMaker Exchange (MME). Patient profiles can be pushed directly to PhenomeCentral from the GPAP PhenoTips instance by the data submitters. The dataset submitter must specifically enable it to be discoverable through MME at the time of submission or in the data management portal. Permission can also be disabled.

The privacy protection legislation requires that each database has its own IDs for the datasets they store.

Submitters (the PI of the group in question) are classed as the data owners and have full rights over the datasets they submit (view, modify, delete). All validated GPAP users have viewing rights over all post-embargo data in the system. They cannot modify data submitted by other users. However, they can “tag” variants as pathogenic or benign.

During the submission process, submitters may choose to restrict data access to members of their own group for a period of up to 6 months. This is designed to allow you and your group priority access to your own data to enable you to analyse it and discover any interesting findings before it becomes visible to other users. During the embargo period, a dataset can be shared specifically with other group/s of users (from another PI/group lead) only by members of the original submitter group.
The embargo period is considered to start at the moment a specific complete dataset (genomic plus phenotypic data) is made accessible to the dataset submitter, which is communicated to the submitter via email. Under certain circumstances, a submitter can request a longer embargo period, but this must be requested at the time of submission and justification must be provided.

No, for ethical reasons, we don’t provide direct access to individual patient data to companies. However, we can help them answer research questions by preparing reports based on aggregated data we store in GPAP, such as “What proportion of patients with disease X have a pathogenic genetic variant in gene XYZ”?

To protect patient privacy, data of individual patients are linked to the unique RD-Connect IDs. Patients’ identity data are not stored on the GPAP and cannot be accessed by GPAP users. Only the researcher, who submitted the data, has the key to identify the IDs corresponding to his/her data. Read more about ethical, legal and social issues.

Yes. The GPAP interface users lets data submitters which other users have accessed a dataset submitted by them and when.

During embargo, datasets are only accessible to the submitter and the users authorised by him/her. If the dataset has been opened to matchmaking, other users performing a matchmaking query may be informed that there is a dataset containing a potential match, but the user cannot see the relevant dataset and must contact the submitter to find out more details or request sharing.
After embargo, datasets are available for searching and querying by other authorised users.

Datasets stored at GPAP are sensitive data and therefore direct download of full datasets is not possible at any time. Download of search results is currently restricted to a maximum of 1000 variants and the option will only be available to the user group that submitted the dataset.

We take data security very seriously and the data in GPAP meet high quality and security standards. To safeguard patient data and protect scientific results, it is an entirely “controlled-access” system. This means that access is strictly regulated, so that only fully validated and authorised researchers can access data within the system, full datasets cannot be downloaded, and every action that a user takes is logged and monitored.
Data is stored in a cluster with a very restricted access policy, limited internet access and daily backups. GPAP security was audited in October 2017 with no major risks being identified. GPAP requests and user actions are safely logged for audit purposes.

Yes, GPAP is compliant with the General Data Protection Regulation, GDPR. (Regulation (EU) 2016/679). Read the Code of Conduct to find out more.

The GPAP doesn’t contain any personally identifiable information of the patients. All data have been pseudonymized, which means they are assigned not to patients’ names but IDs, such as E001234, and can be only identified by the data submitter. Submitters store personal information in the manner appropriate for their own centre.

If a patient wants to remove their personal information from the GPAP database, they should inform the physician who saw the patient originally. The physician will inform the data submitter (if different), who will identify the code for the patient then ask the administrators of GPAP to remove the data associated with that code from the database. The GPAP administrators will confirm to the submitter when this process is complete.

The procedure for becoming a user involves full validation of user credentials including validation of personal ID (passport, ID card) and verification of place of work. Once registered, the RD-Connect GPAP user authentication is performed by a Central Authentication System (CAS) developed by experts from the ELIXIR Spanish node in line with the ELIXIR authentication and authentication and authorization infrastructure (AAI). The CAS has been set up to prevent brute-force attacks and does not store passwords but only password hashes in the user database, and password complexity is enforced. Multi-factor authentication is planned.

At present all sensitive data is housed within its own secure database ecosystem at CNAG and cannot be directly accessed via outside sources or databases but only by authenticated users within the system as already described. Connections to the MatchMaker Exchange API and GA4H Beacon API use secure protocols. Connections are being prepared between the GPAP and the RD-Connect Sample Catalogue via secure means making use of the ELIXIR and BBMRI-ERIC authentication and authorization infrastructures (AAI).

Backups of the database are made daily and the RD-Connect administrators take care of system maintenance and the repairs in case of technical failure. In the highly unlikely event that the system became unavailable in future, users could retrieve all their data to use elsewhere.

GPAP received approval from the relevant Spanish authorities 27th October 2015 (Parc de Salut MAR – Clinical Research Ethics Committee, ref. no. 2015/6456/I). GPAP is included among the personal data files/databases controlled by the CRG (legal entity hosting the CNAG) towards the Spanish law of personal data protection (LOPD).

The underlying database system and its associated intellectual property is owned by the Centro Nacional de Análisis Genómico (CNAG), Barcelona, Spain on behalf of RD-Connect.

Data property remains under the ownership of the submitter (the PI of the submitting group) and (through the PI as intermediary owing to the pseudonymisation) also of the patient. Definitions are included in the Code of Conduct which the submitting PI adheres to by signing an Adherence Agreement.

The oversight of the RD-Connect system as a whole is currently the responsibility of the RD-Connect Community Executive Committee (CEC) and Data Access Committee (DAC).

At the moment, we cannot accept submissions directly from patients with rare diseases. This is because we don’t have the relevant ethics approvals to accept data directly rather than via medical professionals. Furthermore, the GPAP is designed to be used by medical professionals and researchers, and we don’t have the staff to be able to provide you with enough support to interpret your own data. However, we understand that many people with rare diseases want to be able to participate more proactively in research. If you have had your exome or genome sequenced by a commercial provider or by your healthcare provider, we may be able to accept the data, but the way you need to proceed is by asking your doctor or a researcher interested in your disease to register with RD-Connect and upload your data. They will need to be prepared to learn how to use the system and to explore your data.

For data protection reasons, we never receive any personally identifiable information (name, address, date of birth etc.) connected with the data that we hold. Each dataset is only identified by a code. This means we have no way of knowing whether we hold your data or not. If you think that you were part of a research project that sent data to RD-Connect, or if your doctor is one of our contributors, the best way to find out if your data might be in the RD-Connect GPAP is to speak to your doctor directly.

Unfortunately, we currently have no means of providing patients with direct access to their data in the GPAP. We encourage doctors to discuss the results they find in RD-Connect with their patients, and in some cases interested patients may be able to talk to researchers working in their clinical team who use the RD-Connect system (researchers usually really appreciate the opportunity to talk to interested patients!) but this is very much dependent on the centre that submitted your data, so please talk to your doctor directly to find out what is possible.

We take data security very seriously and we have made the systems as safe as possible. This is because the RD-Connect GPAP contains genomic and clinical data which potentially might help to identify a patient, particularly in the case of ultra-rare diseases. To minimise the risks, we have implemented several security measures and allow access only to registered and verified researchers (See Verification process>>). In addition, we do not store any personally identifiable information (such as name, address, etc.) of the patients in RD-Connect. All data have been pseudonymized, which means they are assigned not to patients’ names but IDs, such as E001234, and can be only deidentified by the researcher or clinician who submitted the data in the first place. Only the submitter holds the key that allows the patient’s identity to be recovered, and they hold this in their own clinical systems. Submitters are responsible for storing the key and personal information in the manner appropriate for their institution. The RD-Connect GPAP system has successfully passed externally-conducted security tests showing that it meets high security standards. Data breach is therefore very unlikely, but (as with any online system anywhere in the world) it cannot be excluded with 100% certainty.

Yes, you can remove your data at any time and without providing a reason. To remove your personal information from the GPAP database, please inform the physician who collected it. The physician will inform the data submitter (if different), who will identify your GPAP code and ask the GPAP administrators to remove the data associated with that code from the database. The GPAP administrators will confirm to the submitter when this process is complete.

Yes! Patients must always be asked to consent to their data being used for research. When a doctor or researcher wants to upload data, we ask them to confirm that the individual whose data they are sending has consented to sharing of their data in a way that complies with the legal regulations in the country where the data is coming from.
In all cases, your doctor should have mentioned to you that your data and your biosample (your blood or tissue sample or the DNA extracted from it) might be shared with other doctors and scientists across the world to find a diagnosis or to do research into your condition. Your doctor might not have explicitly mentioned RD-Connect to you – sometimes “broad consent” is given, which means that patients consent to sharing in general and not only for a particular project. This can be useful for example if a doctor wants to use a patient’s biosamples or data in several different projects because they didn’t find the diagnosis or the answer to the research question in the first project, or because different projects deal with different research questions about the same disease. In this case, the patient is just asked the general question about whether they agree to the different possible uses of their biosamples and data, and not to the specific project name.